Karl

[Arvind Narayanan @random_walker]

To understand and empathize with how workers in many or most fields outside software experience advances in AI capabilities, I propose a little thought experiment. https://t.co/QZG24aRiBe

[John Arnold @johnarnold]

Only way to limit coming AI backlash is to start shifting taxes from labor to compute. The average voter needs to see salient benefits from AI. Today we tax labor > compute (income vs corp tax, depreciation for machines not education, payroll tax, etc). This will have to invert.

[Karl @glovguy]

@getjonwithit Your argument only works if properties like “wetness” are fully characterized by their structural role. But that would be the claim you’d need to ground; it’s not a consequence of the emulation hierarchy.

[Karl @glovguy]

@getjonwithit This is a sincere, well articulated argument for a view I think is mistaken.

I see no reason why “wetness” would be preserved in emulation. Even if there is an analogous property in the emulation, and it plays the same computational role, that doesn’t grant it “reality”.

[Josh Kale @JoshKale]

Everyone’s saying OpenAI got the “same deal” Anthropic was banned for.
Read the fine print. They’re not the same:

On weapons:
Anthropic asked for “no fully autonomous weapons without human oversight” = a human involved in the decision.

OpenAI’s deal says “human responsibility for the use of force” = someone accountable, which can happen after the fact.

Oversight ≠ Responsibility. One requires a human before the trigger. The other requires a name on the paperwork after.

On surveillance:
Dario said explicitly: current law hasn’t caught up with AI. The government can already buy your movement data, browsing history, etc without a warrant. AI can assemble that into a complete picture of your life, at scale. That’s mass surveillance without breaking a single law.

Anthropic wanted protections beyond current law.

OpenAI’s deal says the Pentagon “reflects them in law and policy.” That’s existing law as the safeguard, the exact law Anthropic said is insufficient.

Same words. Different agreements. Read them carefully

[jasmine sun @jasminewsun]

200+ Google and OpenAI staff have signed this petition to share Anthropic's red lines for the Pentagon's use of AI

let's find out if this is a race to the top or the bottom

https://t.co/3qgmaLfM0i

[Anthropic @AnthropicAI]

A statement from Anthropic CEO, Dario Amodei, on our discussions with the Department of War.

https://t.co/rM77LJejuk

[NASA Mars @NASAMars]

By studying samples analyzed by the Curiosity rover, scientists have taken another step toward understanding whether life could have ever existed on Mars.

A new study suggests that non-biological sources cannot fully account for the abundance of organic compounds found in a sample collected by the rover.

Dig into the details: https://t.co/zZKCgKEROr

[Andrej Karpathy @karpathy]

Finding myself going back to RSS/Atom feeds a lot more recently. There's a lot more higher quality longform and a lot less slop intended to provoke. Any product that happens to look a bit different today but that has fundamentally the same incentive structures will eventually converge to the same black hole at the center of gravity well.

We should bring back RSS - it's open, pervasive, hackable.
Download a client, e.g. NetNewsWire (or vibe code one)
Cold start: example of getting off the ground, here is a list of 92 RSS feeds of blogs that were most popular on HN in 2025:
https://t.co/dwAiIjlXet
Works great and you will lose a lot fewer brain cells.

I don't know, something has to change.

[Armand Domalewski @ArmandDoma]

@RadioFreeStPaul @mattyglesias Land value tax is distinct from property tax, because it doesn’t tax you on any improvements you make to the land, just on the value of the land itself—this provides a strong incentive to improve the land!

[Karl @glovguy]

@cremieuxrecueil It’s the feed. The feed, and how it’s configured, determines the secondary effects social media has on mental health.

[Omar Khattab @lateinteraction]

crazy that they called it context window when attention span was right there

[cargo short dad @cargoshortdad64]

> order a cold brew
> “is this truly a cold brew or is it iced long black”
> barista looks confused, doesn’t understand
> “I mean was it brewed cold or is this just a black coffee with ice”
> barista laughs, “it’s good coffee sir”
> take a sip
> it’s an iced long black

[Karl @glovguy]

@Noahpinion Absolutely! I’m so glad you’re talking more about social media!There are ways we can change the business of social media to better align the incentives of its content curation.

Unbundle Social Media https://t.co/7Q7AEEPQzK

[Karl @glovguy]

The most common mistake I see when trying to make design patterns for LLM applications: treating agents as something fundamental. Quit reifying agents! It’s agentic tasks that are fundamental.

[Karl @glovguy]

You didn’t rank Cuomo because of his assault allegations and his mismanagement of Covid in nursing homes.

I didn’t rank Cuomo because he gave Andy Byford the cold shoulder until he quit.

We are not the same.

[Karl @glovguy]

@Noahpinion So glad you’re talking about this more! Current social media platforms have major principal-agent problems. Social media should be the primary focus for political change.

[Noah Smith 🐇🇺🇸🇺🇦🇹🇼 @Noahpinion]

It took us a decade and a half, but we're finally collectively waking up to the fact that social media is the main problem with America

[Karl @glovguy]

Content curation is work that is not directly compensated-for on social media. As a result, that work doesn’t not happen as often, or at the level of quality, that it should.

[Karl @glovguy]

@Erwin_AI MCP is an API. Rest API is a protocol for making calls over the internet to a server. MCP is a protocol for defining semantic interactions. Different use cases.

I like to think of MCP as defining “skills” for a LLM.